Good point! It seems like the font-size() mixin isn't used in our tests either.

The tests for this PR failed for stylus (See https://travis-ci.org/twbs/rfs/jobs/512764679#L509). It seems like this line in the Stylus docs is causing this:

If we wished, we could invoke stripe() as if it were a property:
stripe #fff #000

Which means the font-size property always is overridden by the font-size mixin. We could probably fix this by using unquote('font-size') but will still cause the mixin to trigger if the font-size property is used in any custom css.

So we got 2 options for Stylus:

• Add the Stylus mixin anyway and add a warning to the readme. This could break existing implementations though. I'm not sure to what extent since this is the first time this is reported.
• Don't add the mixin and change it's implementation in the readme.

For less everything looks ok. We still need to add some tests, but the implementation depends on the direction we take with Stylus.

Which means the font-size property always is overridden by the font-size mixin.

That is a bummer. I don't think overwriting the font-size property is a good idea, one of the main reasons I was eager to adopt rfs is because you can adopt it "incrementally", that is, without breaking fonts that you've previously made responsive manually, for example. But there may be other things to consider.

In the meantime, I've removed the font-size mixin from the Stylus file and I've written the tests for the font-size mixin. I don't know if that is the best way to write them, what do you think?